The Org Chart of 2040 Will Map Decisions, Not Job Titles

In 2040, you may not open a company’s org chart and see boxes connected by lines. You may see a map of authority instead: which human or AI agent can approve a refund, reorder inventory, change a price, schedule a worker, or escalate a customer, and under what conditions.

That future is already arriving. At Andon Market, a San Francisco retail boutique, an AI agent called Luna directs activities including product selection, pricing, hiring, and scheduling, while human employees operate the physical shop. At Meta, meanwhile, reporting on an AI-driven organizational overhaul has surfaced a very different warning: a sharp rise in code changes to internal software platforms and infrastructure, but a much smaller increase in new or upgraded features reaching users, alongside increases in major technical and security incidents and the employee time spent responding to them.

One is a small shop experimenting with an AI boss. The other is a technology giant redesigning how thousands of people work.

Different scale. Same lesson: AI can accelerate execution long before an organization has figured out how to govern it. 

The moment an agent can take action, the question stops being, “Is the output correct?” It becomes, “Who is accountable for the decision path?” That is the org-chart problem of 2040.

When AI acts, authority becomes the product

Traditional org charts assume decisions move up a hierarchy for approval. That model was built for a world in which humans were the only meaningful actors.

Agentic AI changes the equation. An agent can recommend a price and then change it. It can flag an invoice and then hold payment. It can schedule a worker, issue a refund, or reorder inventory. The distance between recommendation and action is collapsing.

Andon Market makes that shift unusually visible. Its AI manager, Luna, directs activities including product selection, pricing, hiring and scheduling while human employees operate the physical shop, and the emerging stories offer some caution about definitions. The question shifts from whether an AI boss is good or bad to where its authority lives, how that authority is bounded, and who owns the consequences.

AI doesn’t eliminate management problems, though it most certainly exposes them. The organizations that thrive will therefore need to make authority explicit. We see three shifts that matter.

Move 1: Address the organizational structure issues (Minimal Structure)

The future organization may be flatter, but it will not be structureless. Minimal Structure means replacing bureaucracy with lightweight scaffolding: fewer unnecessary layers, but clearer decision rights, interfaces, and escalation routes. Minimal Structure matters because AI can increase activity without increasing outcomes.

Reporting on Meta’s Project OT transformation described a sharp increase in code changes to internal software platforms and infrastructure alongside a much smaller increase in new or upgraded features reaching users, with major technical and security incidents and employee time spent responding to those incidents also reportedly rising. Whatever the ultimate causes, the pattern offers a warning for every company deploying agents: increasing the volume of work is not the same as increasing organizational performance.

Minimal Structure standardizes only what must be standardized:

• Decision rights: who can approve, challenge, override, pause and escalate
• Risk envelopes: what an agent can automate and what requires human sign-off
• Interfaces: how teams and agents hand work off without ambiguity
• Rituals: retrospectives, incident reviews and governance touchpoints that keep learning alive

The distinction is important: Minimal Structure is clarity-first, not headcount-first.

Move 2: Change where leaders focus

When machines take on more execution, leaders have to change what they themselves do. In 2040, a leader’s unit of work will be less about making every important decision and more about designing the conditions under which good decisions happen repeatedly.

We see this shift as a “leadership delta”. That means creating system space for people to act, making the purpose of the transformation understandable, protecting dissent and making accountability visible.

The wider AI reorganization generated a related human warning. WIRED reported dissatisfaction among employees moved into Meta’s Applied AI division, including concerns about whether their expertise would be valued, their opportunities for career progression and the stability of the organization around them. Meta CTO Andrew Bosworth acknowledged that the company had done an ‘atrocious job’ explaining the vision, how employees and their careers would be supported through the shift, and how the organization would evolve.

No company can “AI-native” its way around human meaning.

For leaders, the practical checklist is straightforward:

• Explain the why and the how in plain language
• Make growth opportunities explicit so employees know how their roles can evolve by understanding the work and skills that get rewarded
• Protect escalation and dissent so someone can say “pause the agent” when its behavior is wrong or risky
• Make accountability named and visible (no anonymous ownership)

Leaders in the agentic era aren’t there to be heroic; they’re there to stop authority from becoming accidental.

Move 3: Become more deliberate about connecting AI (Synthetic Symbiosis)

The final shift is from thinking about AI as a tool to designing how humans and agents work together.

We call this Synthetic Symbiosis: a deliberate connection between synthetic execution and human judgment, accountability, and learning.

Andon Market illustrates the question in miniature. Humans still perform physical work while an agent coordinates managerial tasks. The important test isn’t whether the agent is “in charge.” It’s whether the system preserves three things:

• Visible agency: Who made the decision?
• Accountable ownership: Who answers for the outcome?
• Learning: How does the organization get smarter from what happened?

Meta offers a different warning. When synthetic output increases faster than the organization’s ability to evaluate, integrate, and learn from that output, the result can be more remediation rather than more value.

Synthetic Symbiosis isn’t simply “humans plus AI.” It is the design of the connection between judgment and execution. For a 40-person company, that might mean a customer-service agent issuing refunds, a finance agent flagging invoices, a sales agent qualifying leads, or a scheduling agent changing shifts.

The technology is increasingly accessible. The organizational design is the differentiator.

The org chart of 2040: from reporting lines to decision networks

So what does the future org chart actually look like?

Think in layers rather than boxes:

1. Mission: What are we trying to accomplish, and what will we refuse to optimize?
2. Decision: Who can decide, override, pause, and escalate?
3. Teams: How do people assemble around outcomes rather than static functions?
4. Agents: Which agents have permission to act, and within which workflows?
5. Learning: How are decisions, overrides, incidents, and outcomes captured and fed back into the system?

Notice what’s missing: a simple reporting line. The org chart becomes a decision system, one that remains clear even when execution happens at machine speed. At machine speed, semantic drift quickly becomes mission drift.

Five steps founders can start this quarter

Don’t begin with a company-wide reorganization. Start with one controlled experiment.

1. Pick one workflow. Choose a process where an agent could meaningfully reduce cycle time: refunds, invoice matching, lead qualification or customer triage.

2. Write a decision charter. Specify who (or what) can decide, override, pause and escalate, and what must be logged.

3. Define the risk envelope. State what the agent may do, what it must never do, and which conditions trigger human review.

4. Install a learning loop. Capture recommendations, overrides, outcomes, and incidents. Review them regularly.

5. Train leaders for the new role. Develop the skills to create clarity, psychological safety, accountability, and effective human-agent collaboration.

Steps two through four are where the real work happens. They turn an AI experiment into an operating model.

The takeaway

In 2040, the winning org chart won’t necessarily be flatter. It will be clearer: clear authority, clear boundaries, clear accountability, and fast learning.

The companies that get this right won’t simply deploy more AI agents. They’ll know where those agents can act, where humans must judge, and how the organization learns when either gets it wrong.

When knowledge can be quickly synthesized and distilled, wisdom in the form of judgment and discernment becomes the moat.

Because the future of organizational design is not really about AI. It is about authority. And authority remains the hidden architecture of performance.